International Comparison of Contributions to

International Comparison of Contributions to Psychosocial Research on Survivors of Childhood Cancer: Past and Future Considerations Bob F. Last,1 PHD, Martha A. Grootenhuis,1 PHD, and Christine Eiser,2 PHD 1

Pediatric Psychosocial Department, Academic Medical Center, University of Amsterdam, and

2

CR-UK Child and Family Research Group, Department of Psychology, University of Sheffield

Key words childhood cancer; pediatric psychology; review; international comparison; psychological consequences.

Survival rates of children with cancer have increased dramatically over the last decades. Today approximately two thirds of children with cancer survive their illness. In the developed countries, about 1 in every 1,000 adults reaching the age of 20 will be a long-term survivor. This progress in medical care has changed the focus of research in pediatric psycho-oncology from descriptions of the impact of an impending death of the child to the impact on the quality of life (QoL) of the child and the family during and after treatment. In developed Western countries, the advances in medical treatment are a result of an organized effort in study groups (e.g., the Children’s Oncology Group in the United States, the United Kingdom Children’s Cancer Study Group, the Societé Internationale d’Oncologie Pediatrique [SIOP]), as well as of centralized treatment in hospitals specializing in pediatric oncology. These

efforts have resulted in considerable similarity in treatment protocols and survival rates in different countries. However, increased survival raises many new questions, including: What are the costs in terms of late effects on physical, cognitive, social, and emotional functioning? What are the chances of infertility? To what extent will participation in society be hindered by a history of cancer treatment? In addition, ethical questions are raised regarding the need to inform the parents and the children about (possible) late effects. Further questions can be asked about the kind of supportive care needed and how this care can be offered. In this respect, the documents published by the international working committee of SIOP are of interest. These documents offer guidelines to the clinician and coworkers to meet the demands arising from the rights of the child and the parents in obtaining optimal medical and psychosocial care. This committee

All correspondence should be sent to B. F. Last, Pediatric Psychosocial Department, Emma Children’s Hospital AMC, University of Amsterdam, P.O. Box 22700, 1100 DE Amsterdam, Netherlands. E-mail: [email protected]. Journal of Pediatric Psychology () pp. –,  doi:./jpepsy/jsi Journal of Pediatric Psychology vol.  no.  © Society of Pediatric Psychology ; all rights reserved.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on June 12, 2013

Objective To compare the methods and designs, constructs, publication rate, and location of studies in the past 15 years of pediatric psycho-oncology research, across geographic areas and over time. Methods Three recently published reviews on psychological consequences of surviving childhood cancer were reanalyzed according to (1) inclusion of young or older survivors, (2) 5-year periods of publication, (3) country/area (North America, Europe, or Australia), (4) journal type, (5) number of survivors, (6) design of the study, (7) inclusion of a control group, (8) diagnosis, and (9) the use of standardized instruments on different domains. Results The comparisons show some descriptive differences depending on the origin of study, as well as difference over time. Over the years an increase in research from Europe/ Australia over that from North America is shown. In most studies more constructs are measured during the more recent publication period. Conclusions If we are to conduct exemplary work in this area, it is vital to establish national and international organizations for psychologists to parallel those already established for physicians. More attention should be focused on development of interventions for survivors in the future.



Last, Grootenhuis, and Eiser

Evaluation of the Reviews In order to gain insight into the similarities and differences among studies, we decided to reexamine the data presented in the aforementioned reviews. In total, Stam et al. (2001) included 45 articles about 39 studies; Langeveld et al. (2002) included 31 articles about 30 studies; and Eiser, Hill, and Vance (2000) included 20 articles. The review by Stam et al. (2001) was carried out in the following way. Studies published in English since 1985 in journals and books in the field of social sciences, pediatrics, and nursing were identified using computerbased searches in Medline and PsycInfo. The following keywords were used: cancer/neoplasms, survivors, late effects/longterm effects, adolescence/childhood, psychology/

psychiatry. In addition, references cited in the identified studies were searched for relevant information. The review is particularly focused on social and emotional adjustment. Results about physical, neuropsychological, or intellectual functioning of the cancer survivors are not reported, and studies exclusively dealing with functioning in these domains were excluded. The review is focused on young survivors, age 18 or younger at time of study. All types of cancer were included, and there was no strict criterion about length of survival. Although a 5-year period without treatment can be considered as a criterion for survival of childhood cancer, several investigators also consider children with shorter duration of survival. This is due partly to the different survival perspectives for different diagnoses in childhood cancer. An ideal study sample should include a control group of healthy peers and the use of standardized instruments. Because the application of all these criteria would have resulted in too few studies, the investigators decided to include studies using standardized instruments and having a sample size of at least 20 survivors. All 45 articles from 39 studies found are listed in Appendix I. Langeveld et al. (2002) selected the articles for the review as follows. A literature search of English-language studies for the years 1985–2001 was conducted using the databases of Medline, CINAHL (Cumulative Index to Nursing and Allied Health Literature), Embase, and PsychInfo. Key terms included childhood cancer, long-term survivors, and late effects and were combined with dimensions often included as components of QoL, including psychological/social adjustment, employment/health insurance, schools/learning, and QoL/health status. Relevant articles were then hand-searched for further pertinent references. This review has been conducted according to the methodological criteria for the inclusion of studies in the field of psychosocial pediatric oncology recommended by Eiser, Hill, and Vance (2000). These criteria are (1) wellvalidated and reliable measures, (2) a well-matched control group or comparison with culturally appropriate measurement norms, (3) information about demographics and illness and treatment factors (at least cancer diagnosis and time since diagnosis), (4) respondent rate, and (5) use of appropriate rigorous statistical tests. Additional selection criteria included studies in which survivors were the primary source of QoL information, by means of either interviews or completion of self-report questionnaires. Studies with no more than 20% proxies as the primary source of information were included in this review; 7 studies were based on survivors diagnosed before 20 years of age; and 8 studies were based on survivors at least 5 years after completion of therapy. Initially, the selection of the


has already published guidelines for care of long-term survivors (Masera et al., 1996), which stress that each treatment center should offer a counseling center for serious medical and psychosocial problems. Moreover, the committee recommends preventive as well as supportive programs for long-term survivors with specific adjustment problems. Implementation of these guidelines will be encouraged if they are supported by findings of empirical research. At the time the guidelines for survivors were written, the studies done on late effects of childhood cancer were very limited. Three reviews have recently been published by European authors about psychological consequences of surviving childhood cancer. Stam, Grootenhuis, and Last (2001) selected English-language studies published since 1985 that were particularly focused on social and emotional adjustment of young survivors (18 years of age or younger at the time of study). Langeveld, Stam, Grootenhuis, and Last (2002) focused on young adult survivors (mainly at least 18 years old at the time of investigation) and QoL domains pertinent to them. Eiser, Hill, and Vance (2000) reviewed studies about psychological consequences since 1990. These reviews give us the opportunity to compare the results of current research in pediatric psycho-oncology, because they provide information about the type of cancer, age at evaluation and diagnosis, control group, instruments, and results of relevant studies. In this article, the focus is on the research done in the field of survivors of childhood cancer from 1985 until 2000, and especially on emotional and social adjustment. The question is: What are the differences and similarities in the topics, focus, and methods used in the studies included in these reviews from (1) the different developed countries and (2) over time? Finally, possible directions of further research will be discussed.

Psychosocial Research on Survivors of Childhood Cancer

• global emotional functioning (e.g., General Health Questionnaire, Profile of Mood States, Symptom Checklist); • self-esteem/self-concept (e.g., Rosenberg Self-esteem; Piers-Harris Self-concept scale);

• anxiety (e.g., State-Trait Anxiety Inventory, Revised Children’s Manifest Anxiety Scale); • depression (e.g., Child Depression Rating Scale); • quality of life (e.g., Short Form−12, Health Utility Index, Pediatric Oncology Quality of Life Scale); • family functioning (e.g., Family Adaptability and Cohesion Evaluation Scale); • coping (Nowicki-Strickland Locus of Control; control belief scale; social support scales); • posttraumatic stress (e.g., Impact of Events); • intellectual functioning (IQ tests or educational attainment); and • personality (e.g., Personality Inventory for Children). Additionally, we analyzed whether researchers measured behavioral symptoms (Child Behavior Checklist [CBCL]) in their study. Only standardized questionnaires were categorized, with the exception of educational attainment. Studies that consisted of only interviews (mainly about survival aspects such as problems with life insurance, job discrimination, or sexual issues) were included in the reviews but not categorized. All aforementioned data were entered into an SPSS (Statistical Package for Social Sciences) database. In the case of studies published in more than one article, we counted the inclusion of survivors, number of survivors, design, control group, diagnosis, and instruments used per study. The 5-year period and journal type were counted for number of articles. For all variables among the studies, we compared differences in methods and constructs/instruments for study origin (North America, Europe, Australia) and the 5-year periods of publication (see above), using chi-square tests.

Results In total, 76 articles involving 69 separate studies were analyzed. Of these 69 studies, 48 (70%) were carried out in the United States, 16 (23%) in Europe, 4 (6%) in Australia, and 1 (1%) in Canada. Of the 16 studies carried out in Europe, 5 (31%) were in the United Kingdom, 3 (19%) in the Netherlands, 3 (19%) in Finland, 2 (13%) in Sweden, and 1 each (all 6%) in Israel, Austria, and Norway. Five studies reported data in 2 articles (3 studies from the United States, 1 from Australia, and 1 from Canada), and material from an Australian study was published in 3 journals. For further analysis, we grouped together the studies from North America (United States/Canada) and the studies from Europe,


studies in this review was conducted by two reviewers based on the above-mentioned methodological criteria. However, it was found that in most studies, survivors’ social functioning (e.g., education, employment) was not measured with standardized, well-validated instruments. Because they did not want to exclude the social aspects of survivors’ QoL, the investigators decided to include these studies in the review as well, aware of their methodological limitations. A total of 31 articles of 30 empirical studies were found that met the inclusion criteria (see Appendix II). Additional references have been added to this manuscript, including papers published later (see Appendix III). For Appendix III, we searched Medline and PubMed for references with comparable key words as the two reviewers described before. Articles were not hand-searched for additional references, so the appendix does not presume to be conclusive. Articles in the review by Eiser, Hill, and Vance (2000) were included on the basis of (1) publication since 1990, (2) English language only, and (3) use of standardized measures and statistical tests to compare scores with population norms or matched controls. The review included only four studies that were not covered in the other two reviews. These four studies were excluded based on exclusion criteria used in the other two reviews (Eiser et al., 1997; Kupst et al., 1995; Noll, Bukowski, Davies, Koontz, & Kulkarni, 1993; Stern, Norman, Zevon, 1993). In the reviews headed by Stam and by Langeveld, a table was published that includes summaries of the type of cancer, age at evaluation and diagnosis, and information about the control group, instruments, and results. All the data from these tables have been analyzed by the first and second author, on the basis of inclusion of young (